The effects of soil bulk density, clay content and temperature on soil water content measurement using time-domain reflectometry

Download The effects of soil bulk density, clay content and temperature on soil water content measurement using time-domain reflectometry

Post on 11-Jun-2016




0 download


<ul><li><p>HYDROLOGICAL PROCESSESHydrol. Process. 17, 36013614 (2003)Published online in Wiley InterScience ( DOI: 10.1002/hyp.1358</p><p>The effects of soil bulk density, clay contentand temperature on soil water content measurement</p><p>using time-domain reflectometry</p><p>Yuanshi Gong,1 Qiaohong Cao1 and Zongjia Sun2*1 Department of Soil and Water Sciences, China Agricultural University, #2 Yuan-Ming-Yuan Xi Road, Beijing 100094, Peoples Republic of</p><p>China2 ESI Environmental Sensors Inc., 100-4243 Glanford Avenue, Victoria, BC V8Z 4B9, Canada</p><p>Abstract:Time-domain reflectometry (TDR) is increasingly used for field soil water estimation because the measurement is non-destructive and less affected by soil texture, bulk density and temperature. However, with the increase in instrumentresolution, the influences of soil bulk density and temperature on TDR soil moisture measurements have been reported.The influence is primarily caused by changes in soil and water dielectric permittivity when soil compaction andtemperature varies. The objective of this study is to quantify the influence of soil bulk density and temperature, andto provide the corresponding correction methods. Data collected from sand, sandy loam, loam and clay loam showa linear relationship between the square root of dielectric constant of dry soil and bulk density, and a bulk densitycorrection formula has been developed. The dielectric permittivity of soil solids estimated using this formula is closeto that of oxides of aluminium, silicon, magnesium and calcium. Data collection from sandy loam show a noticeabledecrease in measured soil moisture with increase in temperature when the volumetric soil water content is above030 m3 m3. A temperature-correction equation has been developed, which could provide the corrected soil moisturebased on soil temperature and TDR-measured moisture. The effect of clay content has been detected, but it is notstatistically significant. High clay contents cause the underestimation of soil water content in the low moisture rangeand overestimation of soil water content in the high moisture range. Copyright 2003 John Wiley &amp; Sons, Ltd.</p><p>KEY WORDS time domain reflectometry (TDR); volumetric water content; dielectric permittivity; time delay or traveltime; soil bulk density</p><p>INTRODUCTION</p><p>The accuracy of soil water content measurement using time-domain reflectometry (TDR) depends on (1) theaccuracy of time delay measurement and (2) the calibration used to convert measured time delay to volumetricsoil water content. Many techniques have been developed to improve the accuracy of time delay measurement.For example, the switching diode technique has been employed to obtain an unambiguous time mark in theMoisturePoint TDR soil moisture instrument (Hook et al., 1992). Meanwhile, several calibration equationshave been developed to explore the relationship between time delay and volumetric soil water content (Toppet al., 1980; Ledieu et al., 1986; Roth et al., 1990; Herkelrath et al., 1991). Topp et al. (1980) developeda well-known Universal empirical calibration equation between apparent dielectric permittivity Ka andvolumetric water content v. This universal calibration has been validated by numerous reports for twodecades, when it is applied to general soil conditions. However, with the increase of resolution and accuracyof time delay measurement, a discrepancy between Topp et al.s universal calibration and experimental results</p><p>* Correspondence to: Zongjia Sun, ESI Environmental Sensors Inc., 100-4243 Glanford Avenue, Victoria, BC V8Z 4B9, Canada.E-mail:</p><p>Received 21 May 2002Copyright 2003 John Wiley &amp; Sons, Ltd. Accepted 8 September 2002</p></li><li><p>3602 Y. GONG, Q. CAO AND Z. SUN</p><p>has been reported when the universal relationship of Ka versus was applied to soil with high clay content andsalinity (Dirksen and Dasberg, 1993; Jacobsen and Schjnning, 1993; Dalton, 1992; Wyseure et al., 1997).</p><p>The apparent dielectric constant Ka of a material is determined by measuring the propagating time (timedelay) of an electromagnetic (EM) wave in that material. In practice, an EM wave is sent through the materialof interest along a transmission line (probe) buried in it, and the EM wave is reflected back at the end of thetransmission line. The round-trip time T is then measured. According to Maxwells equation, the velocity v ofan EM wave propagating in a material medium with apparent dielectric permittivity Ka can be calculated thus:</p><p>v D CK05a</p><p>D 2LT</p><p>1</p><p>where C is the velocity of an EM wave in free space and L is the length of the transmission line. Multiplyingby two accounts for a round trip.</p><p>Therefore, the apparent dielectric permittivity Ka is</p><p>Ka D(</p><p>CT</p><p>2L</p><p>)22</p><p>The time delay of an EM wave in air over the distance of 2L is given by</p><p>Ta D 2LC</p><p>3</p><p>Combining Equations (1)(3) we obtainKa D</p><p>(T</p><p>Ta</p><p>)2</p><p>or Ka D TTa 4</p><p>The normalized time T/Ta is linked directly to the dielectric permittivity of the material and it will be referredto as the time delay in the rest of this paper.</p><p>As a porous medium, soil consists of materials in three phases: solid soil particles, liquid soil solution andair in soil. The dielectric permittivity of soil, and in turn the measured time delay T/Ta, is a function of thedielectric permittivity of each of its components and the volume fraction of each component.</p><p>The following linear relation between v and T/Ta has been developed by several researchers (Hook andLivingston, 1996):</p><p>v D T/Ta Ts/TaK05w 1</p><p>5</p><p>where Ts is the travel time in dry soil. Similar to Equation (4), Ts/Ta represents the square root of thedielectric permittivity of dry soil. Kw is the dielectric permittivity of the soil solution. By replacing T/Ta andTs/Ta with Ka and Ks in equation (4) we obtain</p><p>v D K05a K05sK05w 1</p><p>6</p><p>Equation (6) shows a linear relation between K05a and v, with slope of 1/K05w 1 and intercept ofK05s /K</p><p>05w 1.</p><p>Ledieu et al. (1986) reported that the calibration equation between Ka and v could be improved byconsidering soil bulk density, although the effects were relatively small.</p><p>Copyright 2003 John Wiley &amp; Sons, Ltd. Hydrol. Process. 17, 36013614 (2003)</p></li><li><p>TDR MEASUREMENT OF SOIL WATER CONTENT 3603</p><p>Dry soil consists of solid particles and air. The dielectric permittivity of solid soil particles is in the range25 and the dielectric constant of air is near unity. The denser the soil, the greater the volume ratio ofsolid particles to air, and the larger the dielectric permittivity of dry soil Ks. The intercept of Equation (6),K05s /K</p><p>05w 1, increases with increasing soil bulk density.</p><p>For soil with high clay contents, the bound water effects can not be ignored. The water phase in soil canbe subdivided into a free water phase and a bound water phase. Free water, also called bulk water, is able torotate freely following an alternating electrical field. Its dielectric permittivity is around 80 at 20 C, attributedto its high degree of polarization under an external electrical field. In contrast, the bound water phase consistsof water molecules that are bound to the soil surface by adhesive, cohesive and osmotic forces (Hilhorst et al.,2001). The rotation of bound water molecules following an applied electrical field is restricted, resulting inless polarization compared with that of free water, and a low dielectric permittivity. Or and Wraith (1999)obtained the dielectric permittivity of bound water of 6, 10 and 14 by harmonic averaging for a bound-waterregion made up of one, two and three molecular thicknesses, respectively. Sun and Young (2001) obtaineda value of 302 for the distance-weighted average dielectric permittivity of bound water, which consists offour water-molecule layers from the particle surface (first layer) to free water (fourth layer) in Rideau clay.Obviously, the slope of the calibration Equation (5), 1/K05w 1, will be different for soil containing differentamounts of bound water, which is directly related to the clay content. The magnitude of the difference dependson the amount of clay and the clay minerals.</p><p>The electrical conductivity (EC) of clay soil imposes a great impact on soil water content measurementusing TDR (Topp et al., 1980, 2000; Malicki et al., 1994; White et al., 1994; Sun et al., 2000). The soil ECcomes from the electrolytes in soil solution and the electrical charged clay colloid surface. The elevated ECincreases the apparent dielectric permittivity (White et al., 1994; Sun et al., 2000; Topp et al., 2000), actingcounter to that of bound water in TDR soil water content measurement, and making TDR less sensitive tosoil texture.</p><p>The dielectric permittivity of free water Kw is temperature dependent. The change of dielectric permittivityof free water with temperature can be described by the following formula (Weast, 1986):</p><p>Kwater D 7854[1 4579 103t 25 C 119 105t 252 28 108t 253] 7where t is the water temperature in celsius.</p><p>When the temperature drops from 20 C to 5 C, the Kw will increase from 8036 to 8612, resulting in achange of the slope of Equation (5) from 01256 to 01208. It is clear that the temperature effects must betaken into account for an accurate soil water content measurement using TDR.</p><p>This experiment quantitatively studied the effects of soil bulk density, clay content and temperature on soilwater content measurement using TDR, and attempts to provide practical calibrations for an accurate soilwater content measurement.</p><p>MATERIALS AND METHODThe soils used in this study were sand, sandy loam, loam, silt loam and silt clay. Sand was dug from theriverbed of Xiao QingHe, Beijing, China. Sandy loam and loam were taken from the experimental field ofChina Agricultural University, Beijing, China. Silt clay loam (cinnamon soils) and silt clay (cinnamon soils)was taken from suburban Beijing, China. The soil type and texture are presented in Table I. The soil wasair dried and sieved (mesh #10). Soil samples were oven dried at 105 C for 48 h and cooled to 20 C ina desiccating chamber for further use. Soil was packed in a PVC cylinder 10 cm in diameter and 40 cm inlength. The time delay was measured using a TDR soil moisture instrument (MoisturePoint MP-917, E.S.I.Environmental Sensors Inc., Victoria, BC, Canada). The probe used in this experiment consists of two 30 cmrectangular stainless-steel bars (13 cm wide, 032 cm thick) separated by 15 cm with epoxy between them.Switching diodes were mounted at both ends of the probe to improve the accuracy of time delay measurement</p><p>Copyright 2003 John Wiley &amp; Sons, Ltd. Hydrol. Process. 17, 36013614 (2003)</p></li><li><p>3604 Y. GONG, Q. CAO AND Z. SUN</p><p>Table I. Particle size distribution of soils under investigation</p><p>Sampleno.</p><p>Soil type Sand (%)(20005 mm)</p><p>Silt (%)(0050002 mm)</p><p>Clay (%)(</p></li><li><p>TDR MEASUREMENT OF SOIL WATER CONTENT 3605</p><p>Electrical balance</p><p>Computer</p><p>ProbeCable</p><p>MP-917TDR</p><p>Mariottebottle</p><p>PVCpipe </p><p>Distilledwater </p><p>Soil particle</p><p>Figure 1. The experimental setup</p><p>of the cylinder was sealed to prevent water loss by evaporation. The soil columns were then moved to thetemperature chamber. TDR measurement started 6 h later, assuming that the soil temperature had reachedequilibrium.</p><p>Measurement of water content for soil with different clay contentsThree soil samples (loam, silt clay loam and silt clay) were packed into a cylinder with a bulk density</p><p>of 133 g cm3. The water content of the soil column was increased from zero to 045 cm3 cm3 in stepsof 003 cm3 cm3 by adding water from the bottom of the cylinder using a Mariotte bottle. The volumetricwater content was measured using TDR and the gravimetric method at each water content level.</p><p>RESULTS AND DISCUSSION</p><p>The linearity of the relationship between time delay (T/Ta) and volumetric water content for the alternativeprobe</p><p>The probe used in the experiment has an epoxy filling between two rectangular metal bars. It is an alternativeprobe structure compared with the common two or three metal rod TDR probes. The question has been raisedabout whether the time delay measured by this epoxy-filled probe is linearly related to that measured bythe rod probe. If the relationship is not linear, then the epoxy-filled-probe-measured T/Ta cannot be simplyconverted to rod-probe-measured T/Ta using two calibration coefficients A and B (MoisturePoint MP-917Manual), and Equation (5) can not be used for the alternative probe. Studies have shown (Knight et al., 1997)that any probe with its rods completely coated by a material with low permittivity has a relationship thatis not linear. However, if the coating is partial, especially when the coating surrounds less than 30 of therod circumference, then the effect of the coating is not significant. Hook et al. (1992) calibrated the epoxy-filled probe on the basis of a linear relationship between the travel time measured by this probe and by atwo-rod probe with the same length. The linearity of this epoxy-filled probe has been further evaluated byFerre et al. (2000) by using a numerical analysis presented by Knight et al. (1997). They concluded that thisepoxy-filled probe has a highly linear response to the soil relative permittivity, which supports the choice of alinear calibration relationship. The testing conducted in this laboratory also shows a highly linear relationship(R2 D 0987) between measured time delay and volumetric soil water content using this epoxy-filled probe insandy soils (Sun et al., 2000).</p><p>Copyright 2003 John Wiley &amp; Sons, Ltd. Hydrol. Process. 17, 36013614 (2003)</p></li><li><p>3606 Y. GONG, Q. CAO AND Z. SUN</p><p>Soil layering effectThe water content in the soil column was not uniform. The soil column was wetter at the bottom, because</p><p>the water was added from the bottom of the column, and the soil was drier at the top because of evaporationfrom the surface. Topp et al. (1982) measured volumetric water content of layered profiles by TDR. Theyfound that the travel time through different layers of soils is additive. Therefore, the TDR-measured volumetricwater content of a layered soil column should be the weighted average of the actual water content of the layers.If the water content of a soil layer changes abruptly, then a reflection may arise at the boundary between thewet layer and dry layers, possibly resulting in an erroneous interpretation of the graphic waveform. In thisexperiment, the soil water content changes gradually from the top (dry) to bottom (wet), so no additionalreflection was found between the start and end of the probe, the positions of which were precisely determinedby the switching diodes at both ends of the probe. Nadler et al. (1991) measured soil moisture using TDRin two layers (dry/wet, wet/dry) combinations and concluded volumetric water contents were found to beaccurately determined by the TDR method, regardless of soil layering. Young et al. (1997) conducted anupward infiltration method to calibrate TDR system and concluded that this method provides a fast andrepeatable calibration, consistent with conventional calibration.</p><p>Averaging error due to water-content-dependent sensitivityFerre...</p></li></ul>


View more >